Map of COVID-19 variant change over time, with Omicron surge

Описание: Updated version:    • COVID-19 variant map with new Omicron sist...  

This map shows how the SARS-CoV-2 (COVID-19) variant composition changed over time. The observed period spans from Dec 2019 to Dec 31 2021.

The last couple weeks are noisy due to low data availability, and likely do not reflect the true extent of Omicron spread.

The map shows the predominant variant (over 50% reported variant) at the given week for each country (NA = grey). "Original" is the initial Wuhan reference sequence. The bottom plot shows the percent of overall variants reported globally. The side bar shows the estimated variant contribution to the overall case number. Delta aa depicts the average amino acid mutations of all sequences from the original Wuhan reference strain that week.

Methods:
6M+ spike protein sequences were taken from GISAID and aligned to spike references for the original Wuhan strain, Beta, Alpha, Gamma, Delta, and Omicron variants. Sequences were filtered for quality (full length and no ambiguity). Sequences were assigned a variant based on closest variant sequence. Sequences were assigned as Other if the difference fell outside the expected distribution for each variant. Metadata was extracted from the sequence headers. Case data was taken from Joh Hopkins University CSSE.

Notebook will be uploaded to Github in the near future.

Notes:
With 6M+ sequences, sequencing errors and database entry errors cause low levels of noise or miss-characterized sequences. This is almost negligible, but may be more apparent in countries with low sequence reporting. The data is also subject to sampling/reporting bias, e.g, if low-income nations sequence less, their actual circulating variants will not count as much towards the total. There is also a low number of reported sequences for the later portion of Dec 2021, which may underrepresent Omicrons current extent--I hope to update with new data soon. Furthermore, this is only an alignment/comparison of spike proteins--full nucleic acid sequences would be more specific and less prone to chance of sequence mischaracterization.